Thousands of artists — ranging from the late Malaysian film legend P. Ramlee to global pop star Taylor Swift — have been identified within datasets used to train artificial intelligence systems, according to a report published by Yahoo.
The finding, as reported by Yahoo, is raising fresh copyright concerns. The core issue is straightforward: creative works appear to have been swept into the material that AI models learn from, and it is not clear that the artists behind those works agreed to it or were compensated.
AI systems that generate text, images, music, or video are built by ingesting enormous quantities of existing content. When that content includes copyrighted songs, films, and other creative output, questions follow about whether using it for training requires permission, payment, or credit.
The span of names cited by Yahoo — from a decades-old regional film figure to a contemporary international superstar — underscores how wide the net of collected material can be, cutting across eras, countries, and genres.
Why it matters: as AI tools become part of everyday creative and commercial life, the unresolved question of who owns the data behind them could reshape how artists are paid and how the technology is legally allowed to be built.